Steve Edwards <[email protected]> writes: > Won't 200 simultaneous calls result in a lot of 'head thrashing' that > would be avoided by staging the recordings to some form of > non-mechanical storage and then copying the the recording at the > completion of the call?
The extX family of file systems probably will not be clever enough to try to unfragment the writes; they will likely all end up in one long fragmented stream. This is exactly what you want if you do not expect to actually listen to more than a small fraction of the recordings. If you do expect to listen to most of the recordings, copying them after they have been written is a great idea. Especially if you have an off-peak time where the disk is idle anyway, or by placing them on RAM-disk if you can afford to lose some. Frankly I would not worry too much, just use reasonably modern hardware with support for write barriers, enable write caching, use a not-too-clever filesystem, and go RAID-1 on just two disks. You could make a ridiculously fast system with NILFS2 and SSD though... /Benny -- _____________________________________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- New to Asterisk? Join us for a live introductory webinar every Thurs: http://www.asterisk.org/hello asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
